don't buy

Do you mean "behave" in the sense of satisfying their specs? Or having the same unspecified behaviour? (I build lots of one-offs, so to me the difference is significant.)

Besides P. being an unreliable source. Which is a bummer--lots of their stuff, e.g. barrel connectors, are cheap and pretty good. It's not that easy to find V-to-K adapters, for instance.

Cheers,

Phil Hobbs

Mostly the latter. We do stuff like off-centering the bias because we want more swing for pulses. The new ERA5's are very different as far as DC bias point goes, and gain seems to be down anyhow. They used to buy from Stanford/Sirenza, then switched to someone else, and just switched again. ThetaJA keeps changing, so I guess chip sizes or something does too. Their support people are either cagy or ignorant, likely both. I once asked them if their MMICS inverted or not, and they didn't know!

Pity. I characterized the ERA-5 for all sorts of things that aren't on the data sheet, and now it looks to be all useless.

John

that kinda spoils the fun !

personnaly ive only had them explode when ive put them in backwards, ... not something I do very often, although ive read quite a few people mention it here.

Colin =^.^=

I hate that. On the other hand, they're pretty cheap...could you just buy 5 reels at once and use them forever?

Cheers,

Phil Hobbs

Since there is really only a couple transistors and bias resistors in them, can't you guys just roll your own amps? I usually do. Heck, we even make our pizza dough from scratch ;-)

Are you allowed to do that in the US?

martin

We found a distributor that had some old ones in stock, so nabbed a reel. That will keep us out of trouble for a good while. For new designs, we'll go to a prime source, Sirenza or Hittite or W-J maybe.

John

ERA-5 has 20 dB wideband gain, down 3dB at about 4 GHz. And it's stable with any impedances. Other parts go to 10 GHz. For a couple of bucks, they're hard to beat. Interestingly, most are Darlingtons, about the last thing you'd associate with speed.

We do make our own pizza dough. Mo uses a recipe she got from her Sicilian grandmother.

John

"colin" wrote in news:iwRRh.2179$I% snipped-for-privacy@newsfe6-win.ntli.net:

Experienced a couple that blew up after being put in frontwards. (Cheap orange bead types; I suspect they were incorrectly marked.)

--Damon

More likely internally shorted with a tantalum whisker, which caused an internal fire.

They also fail from cumulative damage, especially from high di/dt, often caused from putting them on fast rising supply rails.

it's always wise to mount vertically and not too close to the board.:) I've had to repair a few board's and surrounding components due to those flaming up.

Isn't that one of the reason's they are populer ? low ESR ?

Yes, but exceed that nice little di/dt rating bad enough, and they go off like a mini road flare. A common way to do this to use them as input caps, and then hook up a low impedance power supply, eg a large battery or power supply with large output caps etc, then a switch closure or a power connector being plugged in and then kaboom!

Shorting them out would likely also exceed that rating.

MLCC caps are getting better and larger all the time with no major wear mechanisms, and as John has mentioned, the "new" organic polymer caps work good. With careful design considerations and the lower ESR, the ceramic caps can end up cheaper and smaller with less total capacitance. Sometimes the ESR is so low that you need to put a smallish resistor in series with the cap to add some dampening, otherwise you can get LC resonance, especially when a good ground *AND* power plane is not used.

Sure. But didn't you just write that their properties could change any time without much notice? That's the kind of stuff that makes engineers not sleep at night.

Probably you don't quite need performance all the way up to 4GHz. So if your team would spend some engineering effort and money to design your own amp from regular transistors you'd have a re-usable block you could place over and over again without worrying that someone might switch suppliers. Just test and validate it with several RF transistors. The only downside is that the really good RF transistors come from companies like Infineon. Their distribution methods are IMHO old and crusted. My design-in rate with them fell to an all time low (zero, to be precise) after they weren't able to sell me a partial reel for validation despite my statement that I'd pay whatever it takes.

I did that same thing on a switcher for a client. Company could not deliver production quantities (I guess Rebecca wasn't hired yet...), line stop, crisis was brewing. Kicked out all boutique chips, over to jelly-bean parts. It's been about 12 years now and they never had a whiff of a supply side issue in all those years. It cost the client a bundle for my design work but they definitely amortized that in under two years, so over the long haul that has even increased their profit.

Those are the best recipes. As long as grandma taught her what una punta piccola and stuff like that means in ounces ;-)

Try the kneading hook in an electric drill. Man, that rips. I wasn't allowed to do that anymore after marrying :-(

Only Mini-Circuits, because they keep changing suppliers.

That's the kind of stuff that makes engineers

It would occupy 10x the pcb area and likely wouldn't be as fast.

NEC has good stuff and is dependable. Mouser and Digikey stock a lot of their parts.

John

I sure hope they don't start that with they RF transformers. That would not be cool at all.

Not if you'd migrate from 0805 to 0402, and maybe SC-75 packages ;-)

Good to know. But a single supplier still make me a bit itchy.